The purpose of this proposal is to use an existing database to examine ways in which parenting women's social systems impact their ability to engage in and benefit from treatment. Changes in her social network, her ability to fulfill responsibilities in various roles, and the degree of support she receives from family members and peers, constitute varying degrees of social stress. It is hypothesized that women who experience a high degree of social stress (e.g., family conflict, continued contact with drug-using partners or peers, parenting problems) will have more difficulty focusing their attention on treatment. Specifically, women with high social stress are expected to attend fewer sessions and have poorer relationships with counselors than those under lower social stress conditions. Special emphasis will be placed on issues related to children's level of functioning and parenting responsibilities. Specific aims are to a) define measures of social stress stemming from family relationships, peer relationships, and parenting responsibilities among pregnant and parenting women in substance abuse treatment and b) examine how social stressors impact treatment participation. The proposed analyses will use an existing database to examine the impact of women's social systems on their ability to engage in and profit from treatment. Comprehensive admission and during-treatment data were collected on 147 women and their 229 children as part of a CSAT-funded initiative to provide treatment to women and their dependent children. The participating agency housed 20 families in individual apartments, providing daily treatment services for women and child care for children. Data was collected over a five-year period from April 1996 to September 2000. Analyses will be conducted as follows. First, the psychometric properties of measures reflecting potential family and social stressors will be examined and composite measures will be developed. Next, family and social measures associated with during-treatment outcomes will be identified through univariate analyses. Finally, models representing the relative contribution of various stressors in predicting month 3 outcomes will be developed using multivariate techniques.